Voice Cloning App
voice-cloning-app/Voice-Cloning-AppReviewed by AI
Checked by RepoPicks Radar on
What is Voice Cloning App?
Voice Cloning App builds a text-to-speech voice from audio recordings: it cuts the audio into clips with their transcript, trains a model (Tacotron 2) and then reads any text you type in that voice. It runs as a small web page on your own machine at localhost:5000. On Windows there is an installer, but the published one is the CPU-only build, which does not train locally; for training you need an NVIDIA card with 4 GB or more, a source install, or the authors' Colab notebook. The project is old: the last release is from February 2022.
Voice Cloning App is free and open source (BSD-3-Clause) and can replace ElevenLabs, Respeecher, and Murf.
How easy is it to install Voice Cloning App?
B · installer plus one simple step · Windows
- Windows B
- Linux D
On Windows you download one .exe from the releases page and the app opens in your browser, but it only becomes useful after you build a dataset and train or import a voice, so it is a B. Linux is a D: the install script wants Python 3.6 from an extra PPA, ffmpeg, an NVIDIA driver and CUDA, or you build the Docker image yourself.
Voice Cloning App on GitHub
- Stars
- 1,438on Nov 10, 2026
- Downloads
- 19,0374 this week
- Growth
- +1stars in 6 days
Stars over time
Facts
- Runs on
- Windows, Linux
- You need to install
- Nothing extra
- To build from source
- Python, Docker, Graphics card
- Disk space
- 450 MB–1.3 GB · from the installer
- Accounts or keys
- None
- Cost
- Free
- License
- BSD-3-Clause (Permissive)
- Can replace
- ElevenLabs, Respeecher, Murf, Kits AI
- Last change on GitHub
- Dec 2, 2024
Voice Cloning App risk signals (7)
- Attentionstale releases
Last release on 2022-02-07 (1737 days ago)
automatic check
- AttentionDependencies with known vulnerabilities
Pinned dependencies with known vulnerabilities (OSV): flask: GHSA-68rp-wp8r-4726, GHSA-m2qf-hxjv-5gpq, PYSEC-2023-62; requests: GHSA-9hjg-9r4m-mvj7, GHSA-9wx4-h78v-vm56, GHSA-gc5v-m9x4-r6x2; torch: GHSA-3749-ghw9-m3mg, GHSA-47fc-vmwq-366v, GHSA-53q9-r3pm-6pq6; nltk: GHSA-2jhm-w3mp-jcwr, GHSA-2ww3-fxvq-293j, GHSA-3gq4-3j92-5w49; werkzeug: GHSA-29vq-49wr-vm6x, GHSA-2g68-c3qc-8985, GHSA-87hc-h4r5-73f7; joblib: GHSA-6hrg-qmvc-2xh8, PYSEC-2022-288; python-socketio: GHSA-5w7q-77mv-v69f, GHSA-g8c6-8fjj-2r4m, PYSEC-2026-1854; flask: GHSA-68rp-wp8r-4726, GHSA-m2qf-hxjv-5gpq, PYSEC-2023-62; requests: GHSA-9hjg-9r4m-mvj7, GHSA-9wx4-h78v-vm56, GHSA-gc5v-m9x4-r6x2; torch: GHSA-3749-ghw9-m3mg, GHSA-47fc-vmwq-366v, GHSA-53q9-r3pm-6pq6; nltk: GHSA-2jhm-w3mp-jcwr, GHSA-2ww3-fxvq-293j, GHSA-3gq4-3j92-5w49; werkzeug: GHSA-29vq-49wr-vm6x, GHSA-2g68-c3qc-8985, GHSA-87hc-h4r5-73f7; joblib: GHSA-6hrg-qmvc-2xh8, PYSEC-2022-288; python-socketio: GHSA-5w7q-77mv-v69f, GHSA-g8c6-8fjj-2r4m, PYSEC-2026-1854
automatic check
- Attentionvoice cloning without consent notice
The whole purpose is copying a person's voice (topics: voice-cloning, tts) and the README offers "Automatic dataset generation (with support for subtitles and audiobooks)" plus community notebooks for "Youtube data fetching", with no notice about permission from the person whose voice is used or the rights over the recordings.
AI review
- Attentionend of life python required
install.md and install.sh require Python 3.6 ("Installing Python 3.6" from the deadsnakes PPA), a version that stopped getting security fixes in 2021 and is missing from current Linux distributions.
AI review
- Attentionold web server on local port
The app is a Flask web interface on http://localhost:5000/ built on pinned 2020-2021 packages (Flask==1.1.2, werkzeug==2.0.3, python-socketio==5.5.0); the Docker instructions publish that port with "docker run -d -p 5000:5000", which exposes it to the network.
AI review
- Infowindows build cannot train
install.md: "The CPU Only version supports all features except local training", and the three latest releases only offer the cpuonly executable (Voice-Cloning-App-1.1.1-cpuonly.exe).
AI review
- Infodepends on outside services
README.md points to a "Voice Sharing Hub" on heroku, remote training and synthesis in Google Colab notebooks (some made by community members) and the uberduck.ai service for hosting models.
AI review
Evidence is shown in its original language.
What Radar could not check about Voice Cloning App
- Whether the February 2022 installer still runs on a current Windows.
- How good the cloned voice sounds compared with today's services.
- How many hours of recordings and how much time on a graphics card a decent voice needs.
- Whether the linked Colab notebooks and the voice sharing hub are still online.
- Whether a source install still works, given Python 3.6 and torch 1.9 pins.